GENE C

      This gene is in reverse strand. Both programs predict 2 exons: an initial one, identically predicted in the two programs, and a terminal, which differes a lot between them. Furthermore, genscan predicts a promotor region, but, as has been comented previously, we won't take count of this. We will try to reinforce this initial prediction with the later information we obtain.

      We run a blastp against the two predicted proteins (from genscan and geneid) and obtain the same results: only 36 aminoacids matched with the same score (99%) and with the same protein. As these two proteins had only in common the initial exon, we decide to do the blastp only with this exon, and find a match with an unknown protein (protein for MGC:20197), obtained from a cDNA library preparation. We obtain a 100% of similarity in 36 of the 38 predicted aminoacids in the initial exon.

      Putting this information altogether:

• Initial: It is equally predicted in the two programs, and also reinforced by a human EST begining 85 nucletides later and finishing 1 nucleotide after the predicted exon. In the region they share, they have an homology of 100%. There is also a mouse EST reinforcing this exon. Running a blastp with this exon alone, we obtain 100% of homology with the protein BC010181 for MGC:20197 in 36 of the 38 aminoacids it codes for.
• Internal: With this protein, we run a tblastn and see were it matches with our contig. We find that, as we expected, it reinforces the initial exon, but we also see that it predicts two new exons, both strongly reinforced by the same mouse EST that reinforced the initial exon.
• Terminal: The two predicted terminal exons are very different. The one predicted by geneid isn't reinforced by any EST and, running a blastp with this exon alone, we obtain no significative match.
       The one predicted by genscan doesn't have any homologous human protein, but is reinforced by 2 human ESTs. Both have an important homology but after analizing the regions every EST occupies, we decide to use the one which goes from the nucleotide 39859 to 40853. This is because we want it to extend as much as possible to the upstream region and as little as possible to the other side. Otherwise, considering the last exon predicted by the homologous protein and being reinforced by a mouse EST, we try to study this in the geneid results database. There, we find a better terminal exon, which we will take as our choice.

      This can be visulized in GhostView format.

      So, we can conclude that the initial exon is the predicted by both programs and accept the two exons predicted by the tbalstn with the homologous protein. As terminal exon we finally accept the last exon predicted by the homologous protein.

      Once we have this, as the initial exon is confirmed, we select the upstream region with the ss program (nucleotides 45010 to 47250) and run it against the TRANSFAC server to predict the promoter region. We obtain an output file which we convert to gff format (so that we can convert it into PostScript format and visualize the results with the GhostView) with this comand:

grep "^ V\$.*" gen2geneid.promotor.txt | gawk '{print "AC090948", $1, "prom", $3, $3+length($10), $6, ".",".","gen2geneid"}' > gen2genid.prom.gff

And we visualize the results.